Vision provides critical sensory inputs that guide our routine behaviors; as a result, blindness represents perhaps the most devastating deprivation we can experience. To connect perception to action in this context, complex visual scenes must be efficiently represented in the neural activities of relatively small groups of cells; from these signals, particularly salient cues are extracted, integrated with behavioral goals, and linked to the appropriate responses. These neural processes can be broken down into the individual actions of relatively simple microcircuits, small groups of neurons that perform elementary operations that are widespread in the brain, but which subserve distinct purposes in different contexts. This proposal develops the Drosophila visual system as a model in which the functions of these microcircuits can be dissected at the molecular, cellular and behavioral level, and combines techniques drawn from genetics and systems neuroscience to derive new understanding. This proposal focuses on three circuit interactions that are central features to visual processing. First, one fundamental circuit interaction in the visual system is lateral inhibition, a local interaction that occurs amongst photoreceptors that has been proposed to play a number of distinct roles in information coding. Remarkably, in no context do we know the in vivo functions of this interaction. The first goal of this proposal is to define these functions, ad to understand their molecular and cellular implementation. Second, a number of lines of evidence argue that elements of the peripheral visual system are highly tuned to specific cues, for example, patterns of motion, that are of unique behavioral relevance. The second goal of this project is to define the structure and mechanism of action of one such circuit, and to determine how the output of this circuit guides behavior. Third, the emergence of direction-selectivity in th brain represents a long-standing, paradigmatic neural computation with rich theoretical underpinnings. However, the circuit implementations of these theories are only incompletely understood. The third goal of this proposal is to identify and dissect two microcircuits that produce distinct direction-selective tuning properties. These studies will broadly inform our understanding of retinal function in health and disease. As the development of retinal prostheses that directly stimulate specific circuit elements represents an important treatment possibility for blindness, understanding how these circuits can encode behaviorally-relevant visual information represents an important goal.